<h2>Must know</h2>
<ul>
<li><strong>PHP 8.6 introduces partial function application</strong>: PHP 8.6 is rolling out with Partial Function Application (PFA), a feature that lets you create closures with placeholders, making code cleaner and more flexible. This means you can bind arguments and reorder them without writing verbose inline functions, a significant step for the language. This evolution, building on PHP 8.1’s First Class Callables, allows for more powerful manipulation of PHP’s standard library, as reported by Laravel News. <a href="https://app.daily.dev/posts/NdU9ZcZUX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Salesforce goes all-in on AI coding assistant Cursor</strong>: Salesforce has deployed the Cursor AI coding assistant to all 20,000 of its engineers, achieving over 90% adoption in just a few months. The company reports double-digit gains in developer cycle time, code quality, and throughput, with an 85% reduction in legacy code coverage time. Junior engineers are reportedly understanding codebases faster, while senior engineers are offloading repetitive tasks, freeing them for higher-level work. <a href="https://app.daily.dev/posts/7KqSlBkW9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI coding tools boost productivity but raise quality concerns</strong>: Q1 2026 benchmarks show AI coding tools like Cursor and Claude significantly increasing developer throughput, with Cursor boosting daily individual output by 46%. GitHub Copilot leads in daily adoption at nearly 10%. However, a Carnegie Mellon University study found that while initial productivity jumps 30-40%, it returns to baseline within three months, while static analysis warnings and code complexity rise by 30-40% long-term. This suggests a trade-off between speed and maintainability, with concerns about ‘context collapse’ affecting future AI model training. <a href="https://app.daily.dev/feed-by-ids?id=XkFn8TQqF&amp;id=EZUXhYwRZ" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Laravel News to livestream Laracon EU 2026</strong>: Laravel News will be the exclusive live stream and media partner for Laracon EU 2026, offering free access to all talks and exclusive content during breaks. <a href="https://app.daily.dev/posts/Y6tGAKXUQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>jQuery 4.0 drops old browser support, adds security</strong>: jQuery 4.0, the first major update in a decade, removes support for Internet Explorer 10 and Edge Legacy, while adding security features like Trusted Types and Content Security Policy support. <a href="https://app.daily.dev/posts/uMYmsakGM" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New Ruby framework CMDx uses CERO pattern</strong>: CMDx is a new Ruby framework that organizes business logic with composable command objects, implementing the CERO (Compose, Execute, React, Observe) architectural pattern for type safety, validation, and structured logging. <a href="https://app.daily.dev/posts/LmcMOfqU8" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Go developer survey 2025: satisfaction high, challenges remain</strong>: The 2025 Go Developer Survey shows 91% satisfaction with the language, but developers still struggle with best practices, missing features like enums, and finding trustworthy third-party modules. <a href="https://app.daily.dev/posts/1aHMO98uB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Node.js tightens vulnerability reporting requirements</strong>: Node.js now requires HackerOne vulnerability reporters to have a Signal score of 1.0 or higher due to an overwhelming number of low-quality submissions, aiming to reduce triage time for the security team. <a href="https://app.daily.dev/posts/MwudU18Zl" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>SmartOS: a Type 1 hypervisor for advanced virtualization</strong>: SmartOS, an illumos-based Type 1 hypervisor, offers OS and hardware virtualization with features like ZFS, DTrace, and Zones, running entirely from RAM for efficiency and security. <a href="https://app.daily.dev/feed-by-ids?id=o8QQXN3MF&amp;id=Y5ez7xwcB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenEvidence valued at $12B, challenges AI healthcare rivals</strong>: OpenEvidence, an AI-powered medical information platform for doctors, raised $250 million in Series D funding, reaching a $12 billion valuation and reporting $100 million in annual revenue, positioning itself against OpenAI’s ChatGPT Health and Anthropic’s Claude for Healthcare. <a href="https://app.daily.dev/posts/xYWWekBpG" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Palantir CEO says AI will fix immigration, industry says ‘meh’</strong>: Palantir CEO Alex Karp, speaking at the World Economic Forum, declared that AI will eliminate the need for large-scale immigration in Western economies, except for highly specialized roles. This bold claim comes as Deutsche Bank, Gartner, and PwC all report widespread industry skepticism, with most CEOs seeing ‘zero payoffs’ from their AI investments. So, while Karp is busy predicting a future where AI solves all our labor problems, the rest of the industry is still trying to figure out if AI can even pay for itself. <a href="https://app.daily.dev/posts/3s1vrBGbY" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI world models: the next big thing that isn’t quite here yet</strong>: Apparently, ‘AI world models’ and game engines are ‘design opposites,’ with game engines being all about ‘performance and efficiency’ and AI models being about ‘content creation speed.’ The article then goes on to say the future is ‘hybrid systems’ that combine both. So, basically, we’re being told that two things that are different might eventually become less different. Groundbreaking. No actual products, no benchmarks, just a lot of hand-waving about ‘learned hallucinations’ and ‘new forms of interactive content.’ <a href="https://app.daily.dev/posts/OIj9HQhHA" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ChatGPT UI replication: looks good, breaks fast</strong>: A developer found that ChatGPT can visually replicate Apple demo UIs, but the generated code completely falls apart under performance load because it lacks any understanding of underlying system constraints. It’s like a brilliant artist who can paint a perfect car, but it doesn’t have an engine. The solution? ‘Proper mentoring’ for ChatGPT. Because apparently, our AI overlords just need a good coach to stop them from generating pretty but useless code. <a href="https://app.daily.dev/posts/LQOYGhWvV" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
